Enterprise Model Risk and Governance in Financial Services in APAC

Analytic, statistical and increasingly AI-based models permeate every aspect of any financial institution and capture much of that organization's intellectual property, making model management and governance a strategic priority. Of course, all models simplify reality by making assumptions and emphasizing some performance criteria over others. At one extreme, we have many relatively simple but rapidly evolving scoring models combining different attributes to evaluate customers' propensity to perform some action, like click on a web page, or buy a product. By their nature, these highly dynamic models get quick feedback on their effectiveness and can themselves be scored as part of a statistical experiment or incorporated into an Auto-ML/stepwise regression framework. More complex scoring models have a longer delay before the predicted event – like a loan default or an insurance claim, and this means results are inherently uncertain and require a rigorous statistical process before they can be modified. A final group of models are even more complex, such as ALM, RWA, and stress testing models, have many submodels, many data sources, many assumptions drawn from many stakeholders. “At the heart of any of these models are questions of transparency, explainability and inherent risk.” In this webinar, Steven Claxton from SAS and Dr. Chris Marshall from IDC will explore the range of approaches financial services in Asia use to develop, validate, manage and govern different types of models, discussing some common pitfalls and good practices. What you'll learn - Learn best practices across the region for developing, managing and governing different types of financial models - Discover what Asia Pacific regulators look for in model governance - Find the typical challenges financial services face validating and explaining models to regulators - Future proofing model validation & governance for the AI/ML wave